Neurological Complications In Patients With Infective Endocarditis: Insights From A Tertiary Centre

Background: Neurological complications are common in patients with infective endocarditis (IE). Recent data suggest that neurologic events are a major determinant of prognosis, and that surgery is critical in improving the outcome.Objective: To characterize patients with IE and neurological complications and to determine predictors of embolization to the central nervous system (CNS) and mortality.Methods: Retrospective analysis of patients admitted to a tertiary center with the diagnosis of IE from 2006 to 2016. Statistical significance was defined by a p-value < 0.05.Results: We identified 148 episodes of IE, 20% of which had evidence of CNS embolization. In patients with CNS embolization, 76% presented with ischemic stroke. During follow-up, 35% were submitted to surgery and both in-hospital and one-year mortality were 39%. These patients had longer hospitalizations, but there were no significant differences regarding mortality in patients with and without CNS embolization. The independent predictors of neurological complications were diabetes (p=0.005) and the absence of fever at presentation (p=0.049). Surgery was associated with lower mortality (0 vs. 58%; p=0.003), while patients with septic shock had a poorer prognosis (75 vs. 25%; p=0.014). In multivariate Cox regression, human immunodeficiency virus (HIV) infection was the only independent predictor of in-hospital and 1-year mortality (p=0.011 in both).Conclusions: In this population, embolization to the CNS was common, more often presented as ischemic stroke, and was associated with longer hospitalization, although without significant differences in mortality. In patients with CNS embolization, those submitted to surgery had a good clinical evolution, while patients with septic shock and HIV infection had a worse outcome. These results should be interpreted with caution, taking into consideration that patients with more severe complications or more fragile were probably less often considered for surgery, resulting in selection bias.